RESEARCH
UNDERGRADUATE RESEARCH IN ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING
Gold Experience
I was an active lab member in the Multi-Agent Planning and Learning (MAPLE) lab for approximately two and a half academic years (Spring 2016 to Spring 2019). Mentored by Dr. Marie desJardins, my research in the MAPLE lab focused on reinforcement learning and planning. Reinforcement learning is inspired by control systems and operant conditioning, a major area of research in behavioral psychology. This follows with one of the goals of reverse-engineering the brain of making computers think more like humans. I estimate that I worked approximately 870 hours in total.
​
Over the summer of 2017, I was a Robotics Institute Summer Scholar, where I conducted research in computer vision under the guidance of Dr. Christoph Mertz. I worked approximately 500 hours over the span of the summer. For the summer of 2018, I returned to CMU as a Robotics Institute Summer Scholar, where I was advised by Dr. Katia Sycara. There, I developed a novel framework for reinforcement learning in social environments for domestic service robots. During the fall and winter of 2018, I continued working at CMU with Dr. Sycara, where I helped develop a scalable framework for agents to determine which norms are active in a particular environment that can generalize to previously unseen context.
​
At a high level, my research is connected to my Grand Challenge because it is focuses on reinforcement learning and planning, and understanding how to make intelligent agents can provide useful insights into how humans perform these tasks. In particular, my URA project of integrating ethical reasoning with AMDPs is strongly connected with my Grand Challenge because it involved creating a framework for decision-making that is more familiar to humans. To complete tasks, humans make value judgments every day by weighing the features of the particular task with personal and societal values. For example, when we drive to a hospital in an emergency situation, we may decide to disregard some traffic laws, such as speeding, that we might normally obey while still obeying other laws, such as stopping at red lights. Integrating ethical reasoning and to a framework for hierarchical, integrated planning and learning is one step closer to a framework for making decisions that is similar to that of a human.
​
I plan to continue with reinforcement learning research as a PhD student in Machine Learning at Carnegie Mellon University.
LEARNING OBJECTIVES
HOW MY RESEARCH ALIGNS WITH THE GOALS OF THE PROGRAM
During this experience, I:
Showed integrity by performing research that aligns with my core beliefs and values, completing the ethical conduct of research training, and attending seminars about the ethical implications of research.
Demonstrated perspectivism by considering and integrating the perspectives of all individuals on my research teams.
Exhibited perseverance and realistic vision by setting realistic goals and accomplishing them in a timely manner over the span of multiple semesters.
Demonstrated teamwork by researching collaboratively, as well as mentoring other students and being mentored by professors and graduate students.
Showed flexibility by working on research projects as needed.
​
​In addition, I:
Expressed my ideas in an organized, clear, concise, and accurate manner through regularly participating in lab meetings. Because lab meetings are relatively brief, it is important to convey one's ideas and work quickly without sacrificing clarity or accuracy.
Wrote clearly and effectively in discipline-specific formats and learned to effectively connect multiple ideas and approaches by regularly updating my annotated bibliographies on related papers and contributing to multiple published research papers.
Demonstrated the ability to formulate questions and hypotheses within my discipline by proposing my own project and undertaking multiple separate independent research projects.
Showed understanding of the way practitioners think within the discipline and view the world around them by reading relevant papers, conducting literature reviews, and attending various research talks to better understand the state of the field.
Predicted, recognized, and weighed the risks and benefits of the project for others by learning what the underlying motivations of the project and using that information to determine how I could realistically and meaningfully contribute.
Brought new insights to the problem at hand by incorporating my background knowledge and experience in behavioral psychology and neuroscience, as well as reading about relevant cross-disciplinary research.
Was involved in the scholarly community of the discipline and professional societies by joining ACM and AAAI, volunteering at a conference, attending conference(s), and performing professional service through multiple leadership positions on journal committees.